Nanofabrication and analysis

Both interdigitating nanowire structures and microelectrode pads for later integration into microcontact pads were patterned using electron beam lithography (see Chapter 3.2.1). The substrate consisted of a common glass-slide cut in half, thermally evaporated with a 5 nm thick Cr adhesive layer and a 50 nm thick Au layer. The substrate was additionally coated with a 70 nm thick PMMA 950k layer. Here, PMMA coating allowed selective manipulation of the surface by its sensitivity to electron beams and the nanoelectronic geometry could be directly written in the resist. For structuring, the negative mode of PMMA was applied and the steps can be seen on the following figure describing the fabrication of a sensor geometry with one interdigitating nanowire pair at different magnitudes, imaged by electron beam microscopy (see Figure 35). Firstly, the electron beam was guided in a specific pattern over the resist causing it to crosslink at the respective spots. Here, the dose was set to 7mC/cm² for electrode pads, while the dose for nanowire structures was increased to 27mC/cm². The reason for the application of higher doses for nanostructures can be explained by the proximity effect. As discussed in Chapter 3.2.1, if an electron beam hits the substrate, the electrons are able to penetrate both laterally and horizontally in the resist, thereby pre-exposing adjacent areas. Since the electron beam was guided over the substrate in scanning mode, this effect was greatly effecting the exposure in microstructures by pre-exposing neighboring areas, when it scanned over the pre-defined area. Subsequently, a smaller dose was needed to crosslink the resist. This effect was dramatically reduced in nanostructures, where the effective area, exposed by the electron beam, was greatly reduced. Thus, since the proximity effect was smaller, a higher dose was needed for an efficient hardening of the resist. The proximity effect on the nanowire width will be discussed in the following chapter. In Figure 35 panel A, a SEM image directly after exposure can be seen. The structure was written with respect to previous simulations, namely with a pitch of 1 µm, distance to the opposing microelectrodes of 2 µm, having 30 µm in length and 35 µm in width. In addition, the nanowire diameter was set to 50 nm in the design in order to compensate the proximity effect. By electron treatment of PMMA, both its structure and conductivity slightly changed and could be distinguished from unexposed parts. For a better visualization, the exposed

76

parts were colored (blue). In this, the sensing designs could be efficiently transferred from the in silico design to the substrate. Subsequently, the unexposed resist was removed by soaking the substrate in acetone for 3 minutes. Following, the structure could be easily imaged using SEM by the strong differences in electrical conductance and electron density between gold and PMMA (see Figure 35, panel B). Depending on the applied dose for nanostructures, their width ranged from 75 nm to 250 nm. Now, the structured PMMA layer was used as an etching mask to transfer its structure to the gold layer. Finally, the unprotected gold layer was dry-etched in argon plasma followed by a final removal step of PMMA in oxygen plasma. Since PMMA shows high resistivity against dry-etching methods, the gold layer below crosslinked PMMA was well protected from the argon ions. Doing so, the PMMA structure could be efficiently transferred to the gold layer in a 1:1 fashion (see Figure 35, panel C).

Figure 35: Fabrication of gold nanowire sensing geometries using EBL. (A) The electron beam at a high dose factor was guided in specific pattern on the PMMA layer, thereby crosslinking the PMMA chains. (B) Unexposed PMMA was removed in acetone, while exposed PMMA is resistant to this procedure. The sensor design could be efficiently transferred in the PMMA layer with a nanowire width of 75nm to 250nm. (C) Crosslinked PMMA layer was used as etching layer in argon plasma treatment. By removal of unprotected gold by argon ion etching (RIE), the layer below PMMA was protected allowing a 1:1 pattern transfer from the PMMA layer to the gold layer.

Further, the topography of the fabricated nanowires was evaluated to get an insight into the efficiency of the pattern transfer by dry etching strategy. For this, their profile was examined by both atomic force microscopy (AFM) as well as focused ion beaming (FIB) cutting. On Figure 36, the results for both techniques

77 can be seen. In order to get information about the reproducibility of both EBL technique and pattern transfer, a sensor structure of 18 interdigitating nanowires was used for characterization (see Figure 36, panel A). During FIB cutting, the substrate was completely covered with 50 nm Ti for a correct cutting of the structure without any damage. In this, a large cut over the whole width of the nanowires was performed (see Figure 36, panel A, red dotted line) using an argon

Figure 36: Profile examination of a fabricated nanowire array. (A) Focused Ion Beam (FIB) cutting technique allowed observation of profile appearance by ion-induced direct cutting of the device. Fabricated nanowires revealed nearly rectangular profile with a width of 200nm. (B) Atomic force microscopy (AFM) was used to measure the height of the nanowires. Measurements revealed similar heights of the structures between 50-60 nm.

beam under ultra-high vacuum condition and imaged using a SEM setup (Neon 40, Carl Zeiss AG, Germany). On this chip, profile of the exposed parts of the nanowires revealed a near rectangle cross-section with a width of approximately 200 nm (see Figure 36, panel A, right side). So, the developed pattern transfer allowed homogeneous patterning of the nanowire array, as can be seen on the nearly identical appearance of the individual nanowires. Further, in order to examine the height of the wires, AFM (NanoWizard® 4, JPK Instruments AG, Germany) was used for measuring the profile (see Figure 36, panel B). Here, analysis revealed similar heights of the individual nanowires. Their heights were found to be between 50-60 nm which is slightly higher than the expected thickness when comparing with the evaporated layer thicknesses (Cr: 3 nm, Au: 50 nm). The reason for the variation in the thickness of nanowires was due to small changes

78

in deposited film thicknesses by the thermal evaporation of both metals. In conclusion, nanopatterning of the sensing structure was successfully demonstrated, showing nearly identical nanowire heights and widths without any defects and can be therefore used for prototyping of the sensing device.